When you’re in the middle of a family law case, whether it’s a divorce, dissolution, or a dispute about the allocation of parental rights and responsibilities (often called custody), it is tempting to search for a quick answer on an AI chatbots like ChatGPT, Claude, or Gemini . That is especially true late at night, when your attorney’s office is closed. Those “conversations” may offer short-term relief from uncertainty, but there are good reasons to save your questions for your attorney instead.
Privacy is the first concern. A chatbox is not a confidential space. What you type into AI may not be private. Misinformation is another risk. AI chatbots can give answers that are outdated, don’t reflect the law that would govern, or don’t fit your situation. Most importantly, working through your case with a chatbot can undermine one of the strongest protections you can have: the attorney-client privilege.
The attorney-client privilege protects confidential communication between you and your lawyer made for the purpose of obtaining legal advice. It exists so you can speak candidly and share difficult facts along with favorable ones, without worrying that those conversations will be used against you. Your attorney has a professional duty to keep your confidences and generally can never be compelled to testify about them. The privilege also promotes effective counsel by encouraging clients to be open and honest without the fear of repercussions about what they disclose. That candor is what makes effective representation possible, because your attorney can only advise you well when the attorney has the whole story.
A recent federal court decision addressed what happens to that protection when a client turns to AI. Spoiler alert: In United States v. Heppner the court held that a client’s conversations with a consumer AI chatbot were not protected by the attorney-client privilege, even though the client later shared them with his attorney.
Below, we will look at that specific case, and explain why an AI chatbot is no substitute for counsel, walk through the moments in a family law matter when clients are most tempted to turn to AI for fast information, and discuss how AI can be used without sacrificing your attorney-client privilege.
Key Takeaways
- In Ohio, the attorney-client privilege protects confidential communications between a client and their attorney and R.C. 2317.02(A)(1) generally bars your attorney from testifying about them.
- A recent federal case from New York held that a defendant’s conversations with a consumer AI chatbot, which he used on his own without his lawyer’s direction, were not protected by the attorney-client privilege or the work-product doctrine.
- The court explained that a chatbot is not a lawyer, and the AI platform’s privacy policy permitted it to collect, use, and share user data meant the conversations were not confidential and that the party waived his attorney-client privilege when he voluntarily shared information under those terms.
- While the case wasn’t an Ohio domestic relations case and therefore is not binding in Ohio, the law in this area is still developing, and the same logic and reasoning could apply to parties in family law cases in other states.
- You can still use AI as a tool for general education about the legal process, but keep case-specific facts, identifying details, and strategy questions for your attorney.
AI Chatbots Can Quietly Waive Your Attorney-Client Privilege
While not an Ohio family law case, United States v. Heppner is instructive about how judges could treat information entered by a party into an AI chatbot that is part of a legal dispute. The legal reasoning is based on the attorney-client privilege principles, and it could very well be applied in a similar way if the issue came up in an Ohio domestic relations case.
The Heppner case involved a defendant charged with securities fraud in federal court in New York. After he retained counsel, the defendant used the consumer AI chatbot Claude, on his own, to research and work through possible defenses. He later sent those AI-generated materials to his lawyers. When federal agents searched his home, they found and seized print outs of those chatbot conversations, which the government then sought as evidence. Heppner argued that the chats were protected by his attorney-client privilege because he had retained counsel beforehand, used them to communicate with his lawyers, and the AI chat exchange fell under this umbrella.
The court disagreed and ruled that there was no attorney-client privilege to protect the chat exchanges. Therefore the government was permitted to use them as evidence against Heppner. The court also found that documents generated using an AI platform without the direction of counsel are not protected under the work product doctrine.
The court explained that, because an AI chatbot is not an attorney, there was no reasonable expectation of confidentiality in the exchanges. Essentially, the court said:
- A chatbot is not an attorney. The privilege protects communication between client and licensed counsel.
- The conversations were not confidential. The AI platform’s privacy policy permitted the collection of user “inputs” and “outputs,” to “train” the AI model, and disclosed it to “third parties.” By using the AI platform under those conditions, Heppner waived his privilege, and had no reasonable expectation of privacy, when he voluntarily shared information under those terms.
- Sending the materials to his lawyer did not help. Documents that are not privileged when created do not become privileged just because they are forwarded to an attorney.
Why an AI Chatbot Is No Substitute for an Attorney
Ohio’s attorney-client privilege, codified under Section 2317.02 of the Ohio Revised Code, protects confidential communications between a client and their attorney and prohibits a lawyer from testifying about a client’s communications or a lawyer’s advice to a client. An AI chatbot is not a licensed attorney bound by the rules of confidentiality to the person on the other side of the screen.
The statute also contains a warning that applies directly to AI use: if a client voluntarily reveals the substance of attorney-client communications in a non-privileged context, the attorney can be compelled to testify on whatever was discussed. Pasting your lawyer’s advice into a chatbot to obtain a second opinion also risks that protection.
Accuracy is another problem. A chatbot’s response don’t necessarily reflect Ohio law or local practice. Rather, answers often reflect general legal principles and may even be incorrect. A response that may seem reasonable is no substitute for independent legal counsel who knows Ohio law and the facts of your case.
The legal profession takes this risk seriously. The American Bar Association (ABA) has weighed in on this matter. The ABA has released its Formal Opinion 512, its first ethics opinion on use of AI and warns that using generative AI tools with client-specific information creates confidentiality risks that both lawyers and their clients need to understand. The opinion concludes that a lawyer generally needs a client’s informed consent before entering information into a self-learning AI tool. If lawyers are warned to be this careful about what they type into AI, clients should be at least as careful.
Think of it this way: most consumer AI tools permit the company to store, examine, and/or make use of what users type into the chatbots. These practices undermine any reasonable expectation of confidentiality based on privilege. The American Bar Association raised this same concern in Formal Opinion 512.
The Risk of Waiving Attorney-Client Privilege by Using AI in Your Case
You don’t open an AI chatbot in a divorce to get through a difficult night and picture that what you type will end up in the other side’s hands. But the details “discussed” with chatbots are often exactly the kind of information the other side would want in discovery.
Here are specific scenarios that can compromise your attorney-client privilege in family law cases:
- Asking for legal advice and strategy. Never use an AI chatbot to help you create a strategy for an upcoming family law hearing, negotiations with your spouse, or parenting time dispute.
- Drafting and summarizing. Don’t ask AI to draft or review specific language for your property or child support or spousal support provisions. Don’t ask it to summarize an unfinalized part of your case, such as a mediation session, a settlement offer, or a draft parenting plan. These tasks are all your attorney’s job and those conversations belong with your attorney.
- Cutting and pasting. Never share your lawyer’s emails, letters, text messages, or advice into a chatbot for a second opinion. Voluntarily releasing privileged matter destroys the privilege of confidentiality.
- Uploading financial documents. Don’t upload your spouse’s bank statements, pay stubs, tax returns, or text messages into an AI chatbot to try to find out if they are trying to hide assets. Those documents contain sensitive personal information, and some may be covered by court orders limiting how they can be used. Instead, take those documents to your lawyer.
- Venting. This is perhaps the most difficult to forego in any contentious family law dispute. But resist the urge to seek out information and advice from a chatbot about a spouse’s conduct, including infidelity or financial misconduct, while asking for legal-sounding conclusions. You’ll thank yourself later.
None of these communications are protected the way a conversation with your attorney is. What’s more is that if your case becomes more adversarial than amicable, AI chat logs are a prime target for discovery requests for production, just like other written communications like text messages, social media posts, and emails.
Using AI Without Losing Your Privilege
AI doesn’t have to be completely off-limits during a family law matter. The key is to keep your use general and educational.
Here’s a list of a few best practices to avoid a potential and unintended waiver of your attorney-client privilege and keep your communication with an AI chatbot out of discovery:
- Stick to General Education: Use AI to get general information only, such as how the process of divorce works or the meaning of terms. Never use it as a substitute for your attorney for information about the specifics of your own case. Our firm’s articles, including the vocabulary guide, are good starting points.
- Leave out the specifics: Never provide an AI chatbot with names, account numbers, addresses, or any other identifying information about you, your spouse, your children, or financial accounts. Even facts without names can identify you. Any questions you have that include sensitive information like this are for your attorney only.
- Ask your Attorney first: You can always ask your attorney whether their firm offers clients a secure and confidential way to use AI tools as part of their representation.
- Apply the “Evidence Test”: Think before you type by asking yourself the question “Would I be okay with my spouse’s (or ex-spouse’s) attorney (or the Judge) reading this?” If the answer is no, wait to communicate confidentially with your domestic relations attorney.
You don’t have to forego seeking information from an AI chatbot in divorce. But remember that while it can be a helpful research tool, it is no substitute for a truly privileged, confidential conversation with your legal counsel. And if best practices aren’t followed, it could waive your attorney-client privilege.
Talk to Your Ohio Divorce Attorney Before You Talk to AI
You don’t have to avoid using an AI chatbot in divorce altogether. When used with best practices in mind during your family law case, AI can be a helpful tool for understanding the general legal landscape. But, there is no substitute for understanding the line between an organizational tool and a skilled attorney who can advise, advocate, and protect your case.
The family law attorneys at Melissa Graham-Hurd & Associates have experience helping clients understand what information is (and isn’t) protected as they navigate a divorce or other type of family law matter. Their expertise includes counsel on handling documents and communication during a case and how to use AI tools without creating unnecessary risk.
Located in Green, Ohio, halfway between the Summit County (Akron) and Stark County (Canton) courthouses, Melissa Graham-Hurd & Associates, LLC provides compassionate counsel and trusted representation to clients for a wide range of family law matters. If you have questions about how AI could impact your case, don’t hesitate to contact our firm to schedule your consultation.
This article is for general informational purposes only and is not legal advice. Reading it, or contacting us through this website, does not create an attorney-client relationship. Do not send us confidential information until you have established a formal attorney-client relationship with our firm. Please see our full Disclaimer and Privacy Policy.
